Spatial-temporal dynamics of agricultural drought in the Loess Plateau under a changing environment: Characteristics and potential influencing factors

Understanding the evolution characteristics of regional drought is crucial to the sustainable development of ecology and social economy, especially in typical water-limited area. However, the exploring the spatial-temporal dynamics of agricultural drought in the Loess Plateau (LP) under global warming and regional vegetation restoration project is still insufficient, and its potential causes remain unresolved. To this end, in this study, the standardized soil moisture index (SSMI) was constructed based on the root soil moisture data of the Global Land Data Assimilation System, and their suitability in monitoring agricultural drought in the LP was evaluated. In addition, it is proposed to describe the spatial-temporal dynamics of agricultural drought with the migration of drought centroid, and to further analyze the influencing factors of agricultural drought. Results show that: (1) the response time of agricultural to meteorological drought in the LP is approximately 9 months, mainly with existence of moderate drought and drought coverage nearly 10–25%, and the probability of exceptionally drought in a short time is low; (2) from 1982 to 2015, the spatial distribution of drought frequency in the LP gradually increased from southeast to northwest, and the agricultural drought in the northwest has the characteristics of frequent occurrence, short duration and weak intensity, with no significant trend of longer duration and weaker intensity; (3) among the 7 droughts screened out by the Drought migration model, the migration direction of drought centroids are mainly east–west, concentrated in the middle of the LP. In the early and late stage of drought, drought migration trends to have a longer path and faster speed, while in the middle it is shorter and denser; (4) significant increasing precipitation (p < 0.01) in the LP has not alleviated agricultural drought, whereas significant increasing temperature (p < 0.01) is the direct factor inducing agricultural drought, and the implementation of vegetation restoration project further aggravates the risk of agricultural drought. This study not only provides scientific guidance for agricultural drought early warning, prevention and mitigation of drought losses in the LP from the perspective of time and space, but also offers a valuable reference for understanding the impact of large-scale vegetation restoration projects.